U.S. patent number 6,115,589 [Application Number 08/841,242] was granted by the patent office on 2000-09-05 for speech-operated noise attenuation device (sonad) control system method and apparatus.
This patent grant is currently assigned to Motorola, Inc.. Invention is credited to Enrique Ferrer, Kenneth A. Hansen, Charles R. Ruelke, Kevin B. Traylor, Andrew J. Webster, Rajesh H. Zele.
United States Patent |
6,115,589 |
Ferrer , et al. |
September 5, 2000 |
Speech-operated noise attenuation device (SONAD) control system
method and apparatus
Abstract
A SONAD (110) control system (100) detects a received signal
strength (RSSI) for a radio frequency (RF) signal (102), selects a
threshold transfer function (400-404) in response thereto,
generates a threshold control signal in response to the transfer
function, and utilizes the threshold control signal to select the
SONAD threshold value. During operation, the control system (100)
decreases the attenuation of background noise levels for weak RF
signals.
Inventors: |
Ferrer; Enrique (Miami, FL),
Ruelke; Charles R. (Plantation, FL), Webster; Andrew J.
(Basingstoke, GB), Hansen; Kenneth A. (Round Rock,
TX), Zele; Rajesh H. (Plantation, FL), Traylor; Kevin
B. (Austin, TX) |
Assignee: |
Motorola, Inc. (Schaumburg,
IL)
|
Family
ID: |
25284388 |
Appl.
No.: |
08/841,242 |
Filed: |
April 29, 1997 |
Current U.S.
Class: |
455/249.1;
381/94.1; 455/205; 455/212; 455/506; 704/226 |
Current CPC
Class: |
H03G
3/001 (20130101); H03G 3/341 (20130101); H04B
1/1661 (20130101) |
Current International
Class: |
H03G
3/34 (20060101); H03G 3/00 (20060101); H04B
1/16 (20060101); H04B 001/06 (); H04H 005/00 ();
G10L 021/00 () |
Field of
Search: |
;455/205,570,232.1,506,245.1,249.1,250.1,312,296,501 ;375/345
;381/94.1,13 ;704/233,212 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Eisenzopf; Reinhard J.
Assistant Examiner: Appiah; Charles N.
Attorney, Agent or Firm: Fuller; Andrew S.
Claims
What is claimed is:
1. A method for selecting a threshold value for a speech operated
noise attenuation device (SONAD) comprising the steps of:
detecting a signal to noise ratio (SNR) for a signal;
selecting a transfer function, from a plurality of transfer
functions, in response to the SNR;
generating a control signal in response to the transfer function;
and
utilizing the control signal to select the SONAD threshold value
from among a range of threshold values.
2. A method for adjusting a threshold control signal value for a
speech operated noise attenuation device (SONAD) comprising the
steps of:
detecting a received signal strength (RSSI) for a radio frequency
(RF) signal;
comparing the RSSI to a reference; and
decreasing the threshold control signal value when RSSI is less
than the reference.
3. A speech operated noise attenuation device (SONAD) control
system comprising:
a signal strength detector for detecting a received signal strength
(RSSI) for a radio frequency (RF) signal;
a controller, coupled to the signal strength detector, for
selecting one of a plurality of transfer functions in response to
the RSSI; and
a SONAD, coupled to the controller, for receiving a control signal
that alters the SONAD threshold value in response to the selected
transfer function.
4. The device of claim 3 wherein the SONAD threshold value
decreases as RSSI decreases.
5. The device of claim 3 wherein the SONAD attenuation decreases as
RSSI decreases.
6. A speech operated noise attenuation device (SONAD) control
system comprising:
a signal strength detector for detecting a received signal strength
(RSSI) for a radio frequency (RF) signal;
a controller, coupled to the signal strength detector, for
selecting one of a plurality of transfer functions in response to
the RSSI; and
a SONAD, coupled to the controller, for receiving a control signal
that alters the SONAD threshold value in response to the selected
transfer function;
wherein the plurality of transfer functions are selected from the
group consisting of:
exponential responses; logarithmic responses; linear responses; and
non-linear responses.
Description
TECHNICAL FIELD
This invention generally relates to frequency modulated (FM)
communication systems. The present invention is particularly
directed to a speech-operated noise attenuation device (SONAD) that
attenuates intersyllabic background noise (noise between syllables)
in audio circuits such as those found in the audio output stage of
a radio receiver. Specifically, the invention is capable of
reproducing natural sounding speech at both high and low radio
frequency (RF) signal strength levels.
BACKGROUND
Prior art speech-operated noise attenuation devices (SONAD)
attenuate an audio signal in the absence of voice or speech signals
to suppress the apparent level of background noise between
syllables. These prior art circuits typically provide a degree of
intersyllabic noise attenuation which is related to the prevailing
noise level in the signal. When background noise levels vary, the
noise suppression provided by the prior art SONAD circuits varies
as well. In many audio signals, the level of background noise
varies over a wide range so that the effectiveness of a prior art
SONAD in removing intersyllabic noise will vary significantly. A
SONAD that reduces background noise while accommodating changes in
the background noise level is also known in the art.
Unfortunately, such prior art SONAD circuits produce a very
unnatural sounding speech due to suppression of the background
noise when operated under weak RF signal conditions. This is due to
fast changes in attenuation, which produces a dramatic contrast
between no attenuation and full attenuation. Such behavior is
inevitable, since a single threshold point is used in order to
achieve the desired degree of noise suppression. Other known SONAD
circuits attempt to avoid the aforementioned problems by employing
two attenuators in tandem, each having a different threshold point.
Such a device is taught and suggested in U.S. Pat. No. 4,893,349
entitled "FM Communication System with Improved Response to
Rayleigh-Faded Received Signals," issued Jan. 9, 1990, and assigned
to the assignee of the present application. In accordance with the
teachings of this prior art SONAD, the attenuation changes more
rapidly as the input signal level is reduced. This improved
attenuation characteristic tends to minimize speech distortions,
since only the lower level speech sounds are significantly
attenuated without affecting the higher level speech sounds.
Furthermore, the use of separate threshold points more effectively
reduces the noise between syllables to produce a better sounding
speech than exhibited by prior art SONAD circuits.
Notwithstanding the production of more natural sounding speech
accompanied by the use of the SONAD as taught in U.S. Pat. No.
4,893,349, this approach is still characterized by a fast change in
attenuation which produces a contrast between no attenuation and
full attenuation, in situations such as Rayleigh fading, and is
therefore characterized by the production of speech which is still
somewhat unnatural in its sound and makeup.
A need, therefore, exists to provide an improved audio noise
reduction technique which is capable of reproducing natural
sounding speech at both high RF signal strength levels and at low
RF signal strength levels typically attributable to Rayleigh
fading.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a partial block diagram of an FM receiver incorporating
the intersyllabic noise reduction technique in accordance with the
present invention.
FIG. 2 is a detailed block diagram of the speech operated noise
attenuation device (SONAD) of FIG. 1.
FIG. 3 is a graph illustrating the input/output characteristics of
the SONAD device of FIG. 2.
FIG. 4 is a graph illustrating a set of proposed transfer curves
that may be used to control the threshold of the SONAD of FIG.
2.
FIG. 5 is a flow chart diagram of the steps performed by the
receiver of FIG. 1 in order to practice the noise reduction
technique of the present invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
The features of the present invention, which are believed to be
novel, are set forth with particularity in the appended claims. The
invention, together with further objects and advantages thereof,
may best be understood by reference to the following description,
taken in conjunction with the accompanying drawings, of which like
reference numerals identify like elements.
Referring first to FIG. 1, a partial block diagram of an FM
receiver incorporating an audio signal noise reduction technique
and circuit in accordance with the present invention is shown. As
will be appreciated after review hereof, the device 100 is designed
to minimize low-level background noise under strong RF signal
conditions. During weak RF signal conditions, the device operates
to reduce the attenuation of said signals. Simply stated, the
present invention is a SONAD control system methodology and
apparatus that operates under weak RF signal conditions to reduces
the SONAD threshold level thereby decreasing the background noise
attenuation in order to reproduce more natural sounding speech.
Conversely, during strong RF signal conditions, methodology and
apparatus of the present invention operates to increase the SONAD
threshold level thereby maximizing the background noise
attenuation.
As depicted in FIG. 1, the receiver 100 employs an FM demodulator
104 that receives a frequency modulated (FM) input signal 102 and
employs conventional demodulation techniques to recover the audio
portion of the FM input 102. The recovered audio signal is then
routed to receiver filtering circuitry, such as the de-emphasis
network 108. Device 104 also includes a receive signal strength
indicator (RSSI) capable of deriving a signal strength measurement
corresponding to the received FM input 102. The value 112 of RSSI
once derived is then passed to microcomputer 106. Such devices are
well known in the art and will not therefore be discussed in detail
herein.
In accordance with the current invention, microcomputer 106
operates to select one of a plurality of transfer curves that may
thereafter be used to control or establish a threshold for the
SONAD 110. As will be discussed herein in greater detail, each
transfer curve is uniquely selected in order to present the SONAD
110 with dynamic threshold information that permits threshold
selection or tuning in response to a control signal 114 in order to
adjust to particular environmental conditions and thereby maintain
the audio quality of RF signals influenced by said environmental
conditions. In accordance with the teaching of the present
invention, the transfer curves are based upon or are derived as a
function of the RF signal strength of the FM input signal 102.
While it is suggested that device 106 is a microcomputer, it will
be appreciated by those skilled in the art that other methods are
available for implementing this function, without departing from
the spirit of the present invention. One such alternative is to
employ a translinear circuit or a plurality of translinear circuits
capable of providing any number of transfer function curves.
The overall system architecture of FIG. 1 is designed to minimize
low level background noise during strong RF signal conditions,
while attempting to reproduce natural sounding speech during weak
RF signal conditions. This is accomplished through the use of a
SONAD 110 with unity gain for recovered audio above a predefined
threshold and providing for a 2:1 reduction in voltage levels for
signals falling below said threshold. As will be appreciated, the
SONAD 110 may be incorporated into an Application Specific
Integrated Circuit (ASIC) that is positioned after the standard
receiver audio filtering topology but before the audio power
amplifier 116 and speaker 118. The unity gain threshold of SONAD
110 will be real time programmable through a control voltage 114
derived from microcomputer 106 in response to a RSSI in order to
achieve optimum psycho-acoustic effects. The transfer function H(s)
can be realized by hardware or software, and in the broadest
applications can itself be user selectable dependent on operating
environments (i.e.: fading frequency, user preferences). The unity
gain threshold level of the SONAD 110 is initially set to a
percentage of the received signal maximum system deviation and will
be programmable using a 4 bit (16 level) variable voltage
reference. The threshold programming range will be from
approximately 2 percent (%) to 30 percent (%) of maximum system
deviation.
In effect, the SONAD and SONAD control system of FIG. 1 operates to
decrease the background noise attenuation under weak RF signal
conditions in order to reproduce more natural sounding speech,
while increasing the background noise attenuation under strong RF
signal conditions to reduce the level of background noise.
FIG. 2 is a detailed block diagram of a preferred embodiment of the
SONAD 110 of FIG. 1. As depicted, the SONAD 110 receives an audio
input signal 202 from the de-emphasis filter 108. The audio input
signal 202, also defined as V.sub.in' is rectified by full wave
rectifier 204 and filtered by low pass filter 206 to produce a
direct current (DC) voltage signal V.sub.AVG 208. In accordance
with the preferred embodiment, low pass filter 206 is a 1 pole
filter having a cut-off frequency of approximately 10 hertz
(Hz).
The DC signal V.sub.AVG 208 output from filter 206 is supplied to
comparator 212. As shown, the comparator 212 compares the DC signal
V.sub.AVG 208 to a reference signal V.sub.REF 220. As depicted, the
reference signal V.sub.REF 220 is supplied by a programmable
voltage reference 224. In accordance with the preferred embodiment,
the programmable voltage reference 224 is a 4 bit programmable
voltage reference that receives control signal 114 from
microcomputer 106 of FIG. 1 and adjusts the voltage level of
V.sub.REF 220 in response thereto. V.sub.REF 220 is then compared
to V.sub.AVG 208 by the comparator 212.
During operation, when V.sub.AVG 208 is greater in amplitude than
V.sub.REF 220, the switch control signal 216 operates to close S2.
Under this circumstance, the inputs to voltage controlled
attenuator 222; namely, V.sub.REF 220 and V.sub.CNTL 218 will have
identical amplitudes. Conversely, when V.sub.REF 220 is greater in
amplitude than V.sub.AVG 208, the switch control signal 216
operates to close S1. Under this circumstance, the inputs to
voltage controlled attenuator 222; namely, V.sub.REF 220 and
V.sub.CNTL 218 have differing amplitudes. In response to these
inputs, the voltage controlled attenuator 222 will attenuate the
audio input signal V.sub.in 202 in accordance with the following
relationship. ##EQU1##
Based upon the foregoing, it will be appreciated by those skilled
in the art that in the presence of a speech signal V.sub.AVG 208 is
greater in amplitude than F.sub.REF 220. Under this circumstance,
the inputs to voltage controlled attenuator 222; namely, V.sub.REF
220 and V.sub.CNTL 218 have identical amplitudes and no attenuation
is supplied by the voltage controlled attenuator 222. Conversely,
during pauses in the speech signal, V.sub.REF 220 is greater in
amplitude than V.sub.AVG 208. Under this circumstance, the inputs
to voltage controlled attenuator 222; namely, V.sub.REF 220 and
V.sub.CNTL 218 have differing amplitudes. In response, the voltage
controlled attenuator will operate to attenuate the audio input
V.sub.in 202 to increase the background noise attenuation prior to
delivery to the audio amplifier 116 of FIG. 1.
An illustration of a preferred SONAD 110 characteristic response is
shown in FIG. 3. Reducing the SONAD 110 threshold during weak RF
signal conditions minimizes "audio pumping" generated in the
presence of high demodulator noise. Thus, the tunabililty of the
threshold 300 accommodates noise reduction during strong RF signal
conditions, while mitigating noise pumping effects resulting from a
high SONAD 110 threshold in weak RF signal conditions. It is also
evident that the need for audio processing in weak RF signal
conditions is significantly reduced given that any undesired hum or
tonal products present at the demodulator 104 output will be masked
by discriminator noise.
One area of concern is the selection of the nominal setting of the
SONAD 110 threshold 300. This threshold should be selected to
prevent the unacceptable reduction in audio level that occurs as
the distance between the transmitter microphone (not shown) and the
user increases (lower input
deviation level). This condition can be eliminated if the SONAD 110
threshold 300 is set sufficiently low such that deviations at
increased distances from the microphone result in audio input
voltage levels that are at or above the SONAD 110 threshold 300.
Notwithstanding, the SONAD 110 threshold 300 must also be high
enough to provide for sufficient noise reduction in nominal
operating conditions. Preliminary evaluations have shown that the
unity gain threshold can be set high enough to achieve worst case
hum and noise levels of -50 decibels (dB), while receive audio
overdrive settings of >6 dB provide for sufficient dynamic range
to compensate for any reduction in volume level via volume
control.
A family of proposed transfer curves used in accordance with the
preferred embodiment are shown in FIG. 4. Each curve offers a
unique advantage in tuning the SONAD 110 threshold 300 to adjust to
particular environmental conditions. For example, during nominal
operating conditions where strong RF signals (<-30 dB Quieting)
are encountered, it is anticipated that the response of choice will
be proportional to 20Log(V.sub.in) as represented by curve 400.
This provides for moderate reduction of the threshold for signal
levels <-20 dB Quieting, while quickly reducing the threshold
for very weak signal conditions where discriminator noise can
contribute to audio pumping.
In environments where strong RF signal conditions are prevalent,
but fading is frequently encountered, it is anticipated that the
response of choice will be proportional to Exp(V.sub.in) as
represented by curve 402. This response maximizes the noise
suppression in nominal strong signal conditions, while quickly
reducing the threshold during weak signal conditions to eliminate
marginal pumping effects during a fade. In effect, the device of
the present invention decreases the SONAD threshold i.e., reduces
attenuation as RSSI decreases.
In accordance with the preferred embodiment, the linear response,
characterized by curve 404, and further described as (V.sub.in
=V.sub.out), will be selected as a default condition and used when
the operating environment is stable or unknown. Notwithstanding the
above selections, it will be appreciated by those skilled in the
art that several other arithmetic functions can and may be used to
control the SONAD 110 threshold without departing from the spirit
of the present invention. Such other available alternatives
include, but are not limited to non-linear responses.
FIG. 5 is a high level flow chart diagram of the steps performed by
the receiver of FIG. 1 under the control of a controller like
microcomputer 106 of FIG. 1, in order to practice the noise
reduction technique of the present invention. Commencing at start
block 500 flow proceeds to block 502 where the function select
operation is set to default such that the linear transfer function
404 of FIG. 4 is initially selected. From block 502, flow proceeds
to block 504 where the received signal strength of RF input signal
102 is detected and compared to a reference. At blocks 506-510 a
test is performed to detect a momentary RSSI change of more than 15
decibels (dB) lasting less than 750 milliseconds (mS). If so, it is
assumed at block 512 that the device has encountered an environment
where strong signal conditions are prevalent, but fading is
frequently encountered. In accordance with the preferred
embodiment, the response of choice under such conditions is
proportional to Exp(V.sub.in) as represented by curve 402 of FIG.
4. Assuming RSSI has not changed by more that 15 dB at block 506 or
that it changed by more than 15 dB but for longer than 750 ms, flow
proceeds to blocks 518-522 where a test is performed to determine
whether strong RF signal conditions (<-30 dB Quieting) exist. If
so, at block 524, microcomputer 106 of FIG. 1 will select from
memory (not shown) a transfer response proportional to
20Log(V.sub.in) as represented by curve 400. As will be
appreciated, this response will provide for moderate reduction of
the threshold for signal levels <-30 dB Quieting, while quickly
reducing the threshold for very weak RF signal conditions where
discriminator noise can contribute to audio pumping. Otherwise,
flow will proceed from block 518 to block 528 where the linear
response exhibited by the curve 404 of FIG. 4 is selected for
use.
While the preferred embodiments of the invention have been
illustrated and described, it will be clear that the invention is
not so limited. Numerous modifications, changes, variations,
substitutions and equivalents will occur to those skilled in the
art without departing from the spirit and scope of the present
invention as defined by the appended claims. For example, the SONAD
110 can also be completely bypassed via the use of selectable
T-gate, such as switches S1 and S2 as mentioned in FIG. 2 for full
companding compatibility.
* * * * *